Subject: tRNA, glutamine, localization, mitochondria
Y41D4A.6, C39B5.6 and Y66D12A.7 are annotated as subunits of a mitochondrial transamidase which corrects mischarged glu-trna(gln) as gln-trna(gln) and asp-trna(asn) as asn-trna(asn), although the enzyme that creates the mischarged tRNAs in this pathway is missing. On the other hand, the enzyme that directly charges tRNA(gln) is only cytosolic in all eukaryotes studied (PMID: 19417106), and therefore, the syntehsis of gln-tRNA(gln) in mitochondria is not clear. The same puzzle exists in yeast. However, Frechin et al. (PMID: 19417106) determined for the case of glutamine in yeast, that, cytosolic 6.1.1.17 (enzyme for glu + tRNA(glu) -> glu-tRNA(glu)) does the mischarging of glu-tRNA(gln) in mitochondria after being imported into this organelle. Therefore the mischarging-correction (two-step) pathway is adopted for mitochondrial synthesis of gln-tRNA(gln) in the CEL model, but the mischarging reaction is indicated with an "Unknown" gene, since the mechanism is not fully solved and this unspecific assignment will likely not cause a difference with gene expression or proteomics analyses unless organelle-level resolution is desired.
Subject: arachidonic acid metabolism
Products of arachidonic acid metabolism all seem to be at dead ends except for arachd itself, which is a structural component of the cell membrane. The derivatives of arachd produced by this metabolic pathway are important regulators in mamalian metabolism, most notably during inflammation. Although KEGG map for CEL shows no concrete pathway branch, with the additional genes from PMID: 23496871, it is convincing that at least a subset of these products, e.g. prostaglandins, are being generated, probably to be used as signaling molecules as in mamalian metabolism. However, some pathway gaps are still present. These were filled and key products were generated in the reconstructed pathway. These are drained by demand reactions to simulate their production and usage.Since there is no clear evidence for the use of any of the end products in C. elegans, the pathway is simplified to produce only a selected set of regulators. For example, a part of the pathway produces four precursors for Tetrahydrofuran diols. Only one branch, among the four, was reconstructed to represent this sub-pathway.
Subject: 3.1.2.2, acyl-coA thioesterase, acyl-coA hydrolase
Enzyme 3.1.2.2 (see Biosynthesis of unsaturated fatty acids pathway) is associated with the following genes in the model, which have not been captured by the regular GPR pipeline: K05B2.4 or (W03D8.8 or T05E7.1) or C31H5.6 or (C17C3.1 or C17C3.3 or C37H5.13 or F25E2.3). The first matches with human ACOT1, which is cytoplasmic. The second subset in brackets matches with ACOT4, which is peroxisomal, and so does the third single-gene subset. The final set matches with human ACOT8, which is peroxisomal. However, Mitoprot scoring indicates two of the final set of genes (C17C3.1 and C37H5.3) as very likely mitochondrially targeted. This prediction is ignored in the current model.
Subject: EC 1.14.11.2, EC 1.14.11.4, procollagen, protein modification, biomass
Collagen biopsynthesis is based on PMID: 18050497.Amino acid coefficients in reaction RCC0170 that represents collagen biosynthesis was based on average aminoacid composition of collagen proteins (sqt-1 & dpy-17 & lon-3 & rol-6 & dpy-5 & bli-1 & dpy-9 & sqt-2 & rol-8 & bli-2 & dpy-3 & ram-3 & rol-1 & ram-4 & sqt-3 & dpy-13 & dpy-4 & dpy-7 & dpy-8 & dpy-2 & dpy-10).
Subject: Rtotal, Rtotalcoa, Fatty acid composition
Mixed fatty acids in structural metabolites, such as phosphocholine, are represented by 8 Rtotal groups (Rtotal, Rtotal2,...,Rtotal7). Each group is assigned a fatty acid composition based on the type of biomass components it is involved in. Compositions in Rtotal, Rtotal3, and Rtotal7 are determined by corresponding reactions in the biomass reaction set (named as Fatty acid conversions (Rtotal), Fatty acid conversions (Rtotal3), and Fatty acid conversions (Rtotal7), respectively). The others follow the following rules: Rtotal2=Rtotal6=Rtotal,Rtotal4=Rtotal5=Rtotal3. Thus there are currently three different types of fatty acid compositions represented by {Rtotal, Rtotal2, Rtotal6}, {Rtotal3, Rtotal4, Rtotal5}, and {Rtotal7}. The overall composition of structural components depend on the Rtotal groups included. Phospholipids use Rtotal and Rtotal2 (hence having the composition of the first group above), TAG use Rtotal3, 4, and 5 (hence having the composition of the second group above), and ether-lipids use Rtotal, Rtotal6, and Rtotal7 (an equal mixture of first and third groups above). When the resolution in the experimental measurement of lipids is improved in the future, fatty acid composition represented by each Rtotal group may be adjusted by additional biomass reactions to best-fit the measured composition.
Subject: very long chain fatty acids, beta oxidation, fatty acid elongation
Very long chain fatty acids (n>20) are incorporated with a representative compound (C24:0) that is synthesized by fatty acid elongation in the endoplasmic reticulum. It is represented at a minimal percentage in structure, and beta oxidized partially in peroxisome (to C8; after first being oxidized to stcoa and then to pmtcoa) as with other fatty acids. No mitochondrial beta oxidation is allowed for very long chain fatty acids.
Subject: glycerophospholipid metabolism, mitochondria, biomass, 6.2.1.3
Some glycerophospholipid synthesis reactions that are involved in the synthesis of biomass building blocks like pe have high scores of mitochondrial localization. To use metabolic products of such reactions, a mitochondrial lipid biomass reaction needs to be set just like with proteins (after finding evidence that this really is the reason those enzymes and transporters responsible for phospholipid synthesis in mitochondria are used). Using bacterial biomass compositon for mitochondria (except for the cell wall) and yeast or other eukaryotic composition for the rest of the animal is recommended. However, this approach would not change the predictions mathematically and would therefore complicate the model unnecessarily. Thus, inclusion of mitochondrial lipid biosynthesis and lipid degradation pathways were avoided. All reactions related to mitochondrial lipid biomass assembly are contained in the cytosolic compartment in the current model.The localization of 6.2.1.3 reactions may also be affected by this problem. Although some of the genes for this enzyme have exceptionally high mitochondrial scores, reactions of 6.2.1.3 were all included in cytoplasm only. This is because the fatty acids serving as reactants for this enzyme are not present in mitochondria. One potential source for these fatty acids could be the degradation of mitochondrial lipids, but the synthesis and degradation of these are not localized to mitochondria as stated above.
Subject: acyl coA dehydrogenases, acdh family
Acyl-coA dehydrogenases (ACDH) make an important gene family that play roles in the mitochondrial degradation of fatty acids and branched aminoacids. The enzymes of this family use FAD as a cofactor, but this FAD gets oxidized by an electron transfer flavoprotein (ETF) that in turn delivers the electrons to coenzyme Q of the electron transport chain via ETF dehydrogenase. The family has 9 members (PMID: 14728675). C. elegans genome have fifteen related genes: acdh-1, acdh-2, acdh-3, acdh-4, acdh-5, acdh-6, acdh-7, acdh-8, acdh-9, acdh-10, acdh-11, acdh-12, acdh-13, ivd-1, F54D5.7. Specific acyl-coA targets of these genes are not clear from SACURE analysis except for the latter two genes, leaving thirteen to be re-annotated. Some of these genes were annotated and incorporated into the model in three steps: (1) The annotation of ACDH-related genes was manually curated using information in KEGG, Uniprot, and Wormbase. Ten of the fifteen genes were annotated to determine their target acyl-coA groups as short chain acyl-coA (acdh-1 or acdh-2), short and branched chain acyl-coA (acdh-1, acdh-2, acdh-3), medium chain acyl-coA (acdh-7, acdh-8, acdh-10), very long chain acyl-coA (acdh-12), and isobutyryl-CoA (acdh-9). Thus, only long chain acyl-coA's were not associated with any genes. The following convention was used for chain length for short, medium, and very long chain dehydrogenases (based on activity reports in Metacyc): short, 3-8C; medium, 6-16C, very long, 14-24C. (2) All reactions were incorporated by editing the original reactions entered from KEGG. Edits included modification of gene assignments since KEGG seems to have mistakes and nonsense assignments to this end (e.g., medium chain acyl-coA dehydrogenase (EC 1.3.8.7) is assigned to many reactions without regard to specificity; the bad assignments were confirmed by a search in BRENDA for related enzymes and their substrates), and the conversion of the acceptor from FAD to ETF in certain cases. (3) ETF dehydrogenase reaction was used to transfer the ETF electrons to the electron transport chain.The genes for ETF were added to the ETF-dehydrogenase reaction (RM04433); this is a special case since ETF is also a reactant and product in different oxidation states.Since GPRs related to this gene family is not complete (e.g., there is no GPR for long-chain acyl-coA dehydrogenase, some genes are not assigned to any reaction, etc.) there are clearly missing genes in some of the acyl-coA dehydrogenase reactions.Reaction numbering (related to the second item above): ACDH reactions were associated with best-fit KEGG reactions searched in the following order (from better fit to worse): a reaction using ETF as acceptor, a reaction using general acceptor, a reaction using FAD as acceptor.Finally, acdh-1 was associated as a single gene with the conversion of propanyl-coA to acryloyl-coA based on PMID: 27383050 (this modification was made during the first update).
Subject: histone methylation,demethylation,modification,acetylation
Histone modification pathways represent methylation, demethylation, acetylation and deacetylation of histones. For this group of reactions, formula and charge of histone peptides representing the modified and unmodified residues were carefully evaluated based on MetaCyc and significant changes made compared to information in BIGG. Unmodified and most modified (except for trimethylated form) histones were included in demand reactions as reversible. The demand of trimethylated form was forward direction only. Thus, all modifications can take place when a net influx or demand of modified or unmodified forms are imposed, which would be based on the assumed modification levels of the chromosome. Trimethylated histones can be degraded by Lysine degradation (Carnitine biosynthesis) pathway but this turnover is represented by peptides used in the Carnitine biosynthesis pathway.
Subject: fatty acid beta oxidation, peroxisome, mitochondria
Evidence suggests beta oxidation takes place in both mitochondria and peroxisome in C. elegans (PMID: 11025529; PMID: 19496754) and it seems to overlap with pheromone synthesis in peroxisome (PMID: 19496754). All fatty acids with <=20 Carbons (>20 not involved to date) are subject to mitochondrial beta oxidation. Peroxisomal beta oxidation produces short chain fatty acids (PMID: 19496754) rather than proceeding all the way to acetyl coA (stops at octanoyl coA according to basic biochemistry), consistent with the absence of an enzyme involved in last steps (2.3.1.9) in the peroxisome. In the CEL model, peroxisomal beta oxidation is designed to proceed down to C8; the rest of the pothway for C8 fatty acids is handled in mitochondria, consistent with the carnitine shuttle working best with C8 to C18 (MetaCyc).Beta oxidation was limited to the set of fatty acids in TAG and diet.As for unsaturated fatty acids, two auxiliary enzymes are needed (1.3.1.34 and/or 5.3.3.8). Genes for each enzyme localized to peroxisome and mitochondria were determined based on literature (PMID: 11025529; PMID: 19496754).According to PMID: 16982622, there are two pathways for the oxidation of unsaturated fatty acids with double bonds at even numbered carbons. The one that utilizes 1.3.1.34 was adopted, as there is evidence for the presence of this enzyme in both peroxisome and mitochondria. However, the other pathway that uses a multifunctional enzyme in peroxisome may also be available in C. elegans.Except for palmitoyl coA, which represents the backbone of beta oxidation as in KEGG, merged reactions were used. However, shared compounds of a subset of oxidized fatty acids were used as intermediary fatty acids of the overall pathway so that connections between different fatty acids were established. The direct connection of released electrons to the electron transport chain via electron transfer flavoprotein (ETF) is also shown only for palmitoyl coA. For others, FAD was used as the agent in lumped reactions. A reaction that represents the transfer of electrons from FAD to ETF ensures that the overall mechanism is mathematically the same, except that all details are shown with the palmitoyl coa beta oxidation pathway.
Subject: cyclopropane FA
Two cyclopropane fatty acids obtained from diet (C17 and C19) are processed with a truncated betaoxidation pathway that ends when the cyclopropane group is attached to the third carbon from carboxyl end according to an old study in mammals (PMID: 14279117). The enzymes for cyclopropane fatty acid degradation are unknown (PMID: 14711645) and it is not clear whether they are present in C. elegans.Also, a drain reaction is inserted for the last product (C11).Note that, fa17c9coa and fa19c11coa are part of biomass and the degradation pathway is not needed to consume some of the dietary input.
Subject: fatty acid beta oxidation, branched chain fatty acids
Two branched chain fatty acids (C13iso, C15iso, and C17iso) that are part of phospholipids and TAG are subject to beta oxidation as other fatty acids with the exception that the oxidation stops at ivcoa. This compound is connected to the rest of the metabolism in mitochondria via specific reactions in other pathways.
Subject: carbon fixation pathway, pentose phosphate pathway
Some reactions in GPR were associated with or related to carbon fixation in photosynthetic organisms pathway in KEGG. These were linked to pentose phosphate pathway via metabolite s7p, and therefore, grouped under the pathway named "pentose phosphate pathway (indirect)".
Subject: pathway: protein modification by hydroxylation
Protein modification by hydroxylation is a pathway that was formed to represent hydroxylation of aspartate and asparagine residues in some proteins. An important example protein with a modifiable asn residue is HIF, which is inhibited by FIH (factor inhibiting HIF), an enzyme that can be associated with EC 1.14.11.16, under normal oxygen levels (asparaginyl hydroxylation uses O2 as a reactant). This negatively modulates the activity of HIF, therefore HIF activity is induced when O2 levels go down. See also PMID: 14701857.This pathway was kept separate from Glycosylation pathways that also use asn residues on proteins. Thus, the peptides holding the asn residue in these different pathways are different molecules in the model.Compartmentalization of the pathway seems complicated with multiple organelles and therefore it may need to be supported by multiple transport reactions in a future model addressing other organelle compartmentalizations.
Subject: methylglyoxal pathway
Methlyglyoxal detoxification pathway is present in C. elegans according to PMID: 19675139. Methylglyoxal is formed with elevated levels of glycolysis or fatty acid oxidation and is detoxified by being converted to lactic acid. More common path of formation is from g3p and dhap, which is adopted in the current model. This conversion can take place non-enzymatically. There is an alternative pathway of mthgxl generation as mentioned by PMID: 16037240, where mthgxl is formed from acetone. This has been reported for humans and the main enzyme seems to be a cytochrome P450. Due to the difficulty of annotation with this family of enzymes (i.e., specific functions cannot be assigned to the large set of genes in this family) and to the lack of evidence for acetone formation in C. elegans (during excessive fatty acid beta oxidation), this pathway is not adopted. However, it can be reconstructed by an initial non-enzymatic step from acac to acetone and then cytochrome P450 conversion to mthgxl via acetol.To establish the methylglyoxal detoxification pathway, a mitochondria/cytosol transport for methylglyoxal was introduced, and mthgxl forming reactions were placed in cytosol , in order to (i) avoid a pathway hole where g3p -> mthgxl conversion could not take place (since there is no transport for g3p, unlike with dhap), and (2) be consistent with the fact that mthgxl is a by product of glycolysis. The presence of mthgxl and glyoxylase-1 in C. elegans mitochondria and the damage to mitochondrial proteins by mthgxl have been documented (PMID: 18221415), so detoxification is kept in mitochondria. Probably, mthgxl is also detoxified in the cytosol by another enzyme, but the incorporation of this awaits further evidence.Mthgxl can be excreted by a BIGG transport reaction incorporated in the current model. Thus, simulation of methylglyoxal detoxification may require constraining the flux in this transport reaction.
Subject: fad synthesis, riboflavin metabolism and transport
Riboflavin is a precursor for the synthesis of FAD via FMN. It is obtained from diet in the case of C. elegans as it is abundant in most bacteria. However, E. coli metabolome also shows abundant FAD and FMN. In addition, there are membrane-bound enzymes based on predicted localization which can convert FAD and FMN to riboflavin, with subsequent transport as riboflavin (FAD and FMN are large molecules with no indication of transport from extracellular environment in yeast and human models). Thus, compartmentalization was used to design a system that can utilize all three molecules that can be potentially obtained from the diet.
Subject: glycogen degradation
Glycogen degradation pathway based on MetaCyc and Berg, Tymoczko, and Stryer, 2002 (http://www.ncbi.nlm.nih.gov/books/NBK22413/). Berg et al. textbook suggests that about 90% of glycogen residues are phosphorolytically cleaved and about 10% hydorolytically cleaved. The pathway is designed such that the overall reaction (sum of RCC0043-46) adds up to these proportions; the degradation ratio is arbitrarily set at 0.5 in RCC0043 and the stoichimetry of the rest of the reactions were determined to reach the 1/9 ratio.
Subject: reaction stoichiometry, mass balance, polymers
When there is a polymeric addition involved in a species and is represented by x(mer)n, with x being the rest of the molecule (as in protSfar of RC09845), the breaking up of this polymer as a separate polymeric compound is also represented by x(mer)n (x: the terminal sections of the polymer; e.g., pepd of RC09845) due to the formula conventions, but it actually should be x(mer)n-1, since the terminals are lost in x hence removing one monomer. In case of hydrolysis for polymer degradation, the difference between x and the monomeric unit is H2O (again as in RC09845).
Subject: ubiquinone synthesis, pathway gaps
The initial part of ubiquinone synthesis includes enzymes that are either unknown or uncharacterized. As a first approximation the model uses what is assumed for the human model (PMID: 17267599), which has the same set of restraints. The localization of reactions down to the production of 4hbz is not clear, although there is more evidence for cytosolic localization (for the enzyme EC: 6.2.1.12 and associated genes). The original model arbitrarily placed these reactions in mitochondria together with the rest of the pathway, for which there is higher confidence of mitochondrial localization. During the first model update, reactions down to the production of 4hbz were introduced also to cytosol. Since this was done using a set of reactions, original FBA localization scores that were obtained by one-by-one evaluations are overruled. In addition, 4hbz was made secretable since the pathway after this point seems to be enriched in Gonad, while the pathway before this point is depleted there and enriched in Intestine (mostly based on acs-14 gene expression) according to Cao et al.'s dataset (PMID: 28818938). Overall, the updated reconstruction is consistent with localization evidence and gene expression trends and allows both mitochondrial and cytosolic initiation of the ubiquinone biosynthesis pathway.
Subject: carnitine biosynthesis
Carnitine biosynthesis pathway was extracted from KEGG Lysine degradation pathway and designed according to MetaCyc (PWY-6100) and one of the references therein (PMID: 17944936). The following steps were taken to reconstruct the pathway: 1) KEGG subpathway for Lysine degradation starting from Protein lysine and ending in Carnitine was used as template. 2) Pathway uses specific proteins such as Cytochorome C and histones. A generic peptide was used as a holder of the lysine residue to be transformed into carnitine (pepX). This peptide was triple methylated in a merged reaction of 2.1.1.43. 3) The modified peptidyl lysine was released by a hydrolysis reaction by an uncharacterized enzyme (3.4.-.-). The holder peptide is also recovered in this reaction. 4) In addition to 3.4.-.-, the step carried out by 4.1.2.- was included without gene assignment, as this is also an uncharacterized enzyme. 5) The step associated with 1.14.11.8 was localized to mitochondria according to PMID: 17944936 and the unpublished reports that the associated gene (gbh-2) is located in mitochondria in C. elegans (WormBase ID: WBPaper00011400). The necessary transport reactions were inserted accordingly.
Subject: rhodoquinone, fumarate reductase
A fumarate reductase is assumed to exist in C. elegans based on the following evidence. 1) C. elegans possesses rhodoquinone 9, typical electron carrier used by eukaryotic fumarate reductase systems (PMID: 10545216). 2) Gene F48E8.3 best matches with Ascaris suum fumarate reductase. A. suum is a parasitic nematode that uses its fumarate reductase system in anaerobic conditions (PMID: 7739664). 3) The same gene is orthologous to yeast soluble fumarate reductases (best match is mitochondrial, according to wormbase). 4) The KEGG orthology group that is best associated with this gene is K18561, a novel group for fumarate reductase activity. 5) Under microaerobic conditions, wherein fumarate reductase is expected to be functional, F48E8.3 is up-regulated and rhodoquinone levels are relatively high (PMID: 12875742, PMID: 16522328,PMID: 10545216). Based on all these, the putative fumarate reductase F48E8.3 is incorporated into the model with a reaction that does the reverse of succinate dehydrogenase. The latter was made irreversible. The fumarate reductase reaction was localized to mitochondria due to pathway.In addition, the synthesis of rhodoquinone-9 was represented by a conversion from ubiquinone 9, based on a template reaction in MetaCyc. The actual pathway is known to exist in some eukaryotes, but the exact mechanism is unknown to date.
Subject: Degradation of alkylacylglycerolipids
Glyceryl-ether monooxygenase (1.14.16.5) is an enzyme that can start the degradation of metabolites from ether-lipid metabolism (PAF branch) (PMID: 9436181). The gene associated with it (agmo-1) was shown to be important in host defense against bacteria (WormBase ID: WBPaper00043001), which is consistent with its functions in mammals (PMID: 9436181). The likely substrate for this enzyme in CEL metabolism is ak2lgchol, which is also consistent with reported enzyme specificity (PMID: 9436181). The pathway for degradation of ak2lgchol was reconstructed based on the generic pathway in PMID: 20643956 (see Figure S1 therein). The pathway was named "degradation of alkylacylglycerolipids" and generic reactions for this pathway were spared (R04044, R08372).
Subject: molybdenum cofactor biosynthesis
Molybdenum cofactor biosynthesis pathway was formed mainly using MetaCyc (PWY-6823). Two merged reactions were used for the beginning (RCC0142) and the end (RCC0143) of the pathway. In the first part of the pathway, where various enzymes are modified to carry sulfur atoms, one reaction in MetaCyc was unbalanced (RXN-12473; EC 2.8.1.11). A model provided by PMID: 18650437, which was suggested for the human pathway, was used to make a balanced merged reaction. As it turns out, an electron acceptor was missing; currently, thioredoxin is used as suggested by PMID: 18650437.The enzyme catalyzing the first reaction of this pathway (cysteine desulfurase, 2.8.7.1, or NFS1) is localized to both mitochondria and cytosol in humans (PMID: 9885568). The targeting of the same gene is achieved by alternative splicing which removes the mitochondrial signal at the N terminal for cytosolic localization. A mitochondrial reaction of this enzyme helps the formation of iron-sulfur cluster proteins, although the molybdenum cofactor pathway reaction is cytosolic. Both of these reactions are present in the the CEL model (RMC0132 and RCC0142). However, the mitochondrial targeting sequence could not be detected by Mitoprot (it was detected for human NSF1). Nonetheless, the mitochondrial reaction's GPR is maintained as there seem to be no paralogs for the NSF1 gene. It can be assumed that unknown mechanisms may target this enzyme to both compartments as in humans.To drive molybdenum cofator biosynthesis, a demand reaction and a transport (uptake) reaction were inserted for molybdate. Also, it is included as part of the soluble component of bacterial diet.
Subject: nadh dehydrogenase, ubiquinone
Gene-KO-enzyme-rxn assignments for nadh dehydrogenase (ubiquinone) (R02163) were not appropriate. It was redone manually using the definition of the complex in KEGG in pathway cel00190. The four lines of E (eukaryotic) complexes were connected with AND's in the respective order.Three different versions of R02163 were used: RMC0006 from BIGG human model and RMC0007 and RMC0008 from the yeast model.
Subject: propionate production, transport
C. elegans have been shown to have increased levels of propionate in its exometabolome (PMID: 23029411) when under anaerobic conditions. To make the network able to produce propionate, enzyme 3.1.2.18 was introduced. To carry the product from mitochondria to extracellular environment, mitochondrial and extracellular transport reactions were made reversible, unlike with the BIGG human model (PMID: 17267599) where the original (irreversible) reactions came from. This is consistent with the fact that the new human model (Recon X; PMID: 23455439) indicates that these are diffusion-based transport reactions. Further, Recon X supports reversible extracellular transport of propionic acid when it is coupled with reverse transport of some other organic acids.
Subject: retinol, vitamin A, beta carotene
There are retinol and beta-carotene processing enzymes and expression reports and analysis of beta carotene monooxygenase (BCMO) enzyme in the literature (e.g., PMID: 19556237), which implicate the presence of a retinol pathway in C. elegans.Bcmo-1 and bcmo-2 are the two genes for BMCO in C.e elegans which are not designated in KEGG as such (or as anything else) but annotated in WormBase. Annotation is not clear in myKEGG annotation either. However, orthology with human genes suggest that bmco-1 encodes EC 1.14.99.36, while bmco-2 encodes 3.1.1.64.All related genes are used in CEL network within the following pathway model: beta carotene gets converted to retinal by bmco-1, retinal is reversibly converted to retinol, which is taken by a pathway to retinol trans and cis retinyl esters used for storage. In this sub-pathway, fatty acyl chain of the ester was taken as palmitate, as in KEGG. The human metabolic network model uses Rtotal2, which links this pathway to phosphocholine, but humans seem to have LRAT (2.3.1.135) instead of DRAT in C. elegans (2.3.1.76), so this model was not followed. Retinol can also be converted to 9-cis-retinol, which is not specified in KEGG in detail, but only shown by an arrow. In human model, this conversion is attributed to 5.2.1.7, but without any genetic association. BRENDA also does not show this exact reaction for either of the enzyme numbers. Thus this reaction is incorporated without a specific gene in the CEL model, although bcmo-2 is the likely candidate. 9-cis-retinol is then converted to 9-cis-retinal. Demand reactions are entered for trans retinal and 9-cis retinal, as well as the two retinyl esters (trans and cis). Demand reactions for the fatty acyl ester forms are made reversible as these are the storage forms of Vitamin A. Beta carotene is transported in as it is unlikely to be a part of bacterial diet, at least for E. coli.The source for beta carotene may be plants in the wild and peptone in the lab. Since axenic medium does not contain Vitamin A derivatives, worms are likely to grow in the absence of them.
Subject: bacterial degradation, reaction localization
Bacterial degradation was compartmentalized into extracellular compartment for the release of macro components such as proteins ("Bacterial degradation" reaction), and cytosol for the rest of the degradation pathway except for the soluble component. Although the degradation of most macromolecules most likely takes place in the intestinal lumen, all macromolecules are first transported into the cytoplasmic compartment to avoid unnecessary complexity. This is because the actual compartmentalization of digestion mechanisms is not clear and also because changing the order of transport and degradation (first extracellular degradation followed by transport of building blocks into cytosol) would be equivalent to the current simple model. If intracellular and extracellular degradation mechanisms and specific enzymes for these processes are determined at higher resolution and confidence, then more detailed compartmentalization of digestion reactions and genes can be implemented in future versions.
Subject: thiamine
Thiamine (Th) is an essential vitamin that cannot be produced by animals, hence it has to come from the diet. Its active form is thiamine pyrophosphate (ThPP), the dominant species in any organism. According to PMID: 20492686 total thiamine in E. coli is about 1nmol/mg protein and ~95% of it is ThPP, with the rest mostly being ThMP. The dietary thiamine composition was taken arbitrarily as 90% ThPP and 10% Th when designing the uptake (degradation) reaction.Th is more efficiently transported than Th derivatives so ThPP gets dephosphorylated in the intestine before uptake into cells in mammals (PMID: 15514058, and references therein). According to de Jong et al. (PMID: 15514058), this is likely the case in C. elegans, too, although the mechanism seems unknown. A reaction for dephosphorylation is therefore included with an unknown enzyme.De Jong et al. (PMID: 15514058) also claim that ThPP can be directly transported through plasma membranes, but the mechanism is not clear and the results seem inconclusive. Thus, the current mammalian model with dephosphorylation of ThPP to Th and uptake in the form of Th followed by ThPP formation by the TPK enzyme is used in the current reconstruction. Also, TPK reaction going from Th to ThPP seems thermodynamically reversible. This may explain de Jong et al.'s observations since Th-->ThPP converison of TPK is coupled to Th transport (perhaps, dephosphorylation is also involved in the same mechanism; and note that de Jong et al.'s observations are based on a partial loss of function of TPK by mutation). It seems difficult to disentangle phosphorylation, dephosphorylation and transport of thiamine with indirect experiments.ThPP is also used in mitochondria. Based on PMID: 12014993, mitochondrial uptake happens in mammals, but the mechanism is not known. Thus an artificial transport equation is included. BIGG human model (PMID: 17267599) offers water or hydroxide exchange for this transport, but acknowledges that the actual mechanism is unknown based on the same reference.Thiamine can be uptaken without direct energy expenditure but Na/K ATPase activity is required for its export (see discussions in PMID: 15514058 or PMID: 10964259). This seems to be the case for the redistribution of thiamine in the worm, according to PMID: 15514058. Thus, ATPase connection may be more relevant when converting the generic model to a compartmentalized animal model.
Subject: vitamin B6
Animals including C. elegans cannot synthesize vitamin B6, so they are dependent on dietary forms via a salvage pathway (PMID: 15483325). Vitamin B6 metabolism reaction set from KEGG was reduced to the vitamin B6 salvage pathway in MetaCyc. This pathway represents the conversion of three forms of vitamin B6 that can be absorbed from the environment: pyridoxine, pyridoxal, and pyridoxamine. They all end up in pyridoxal phosphate, which is the main coenzyme used in metabolism. Although this compound is consumed or produced in a few reactions, the general usage is not well represented this way; therefore a demand reaction is also inserted. The flux of this demand reaction should not exceed what is to come from the diet. The input of the three Vitamin B6 components above was connected to the soluble pool in bacterial diet. Coefficients were determined based on ~0.3nmole/mg cell total vitamin B6 according to PMID: 4945178.
Subject: vitamin B12, cobalamin
Vitamin B12 cannot be synthesized de novo by C. elegans. Nor can it be synthesized by E. coli (PMID: 23772381). But both can convert cobalamin (basic vitamin B12; different forms are available depending on the moiety linked to Co within the porphyrin structure) to the active coenzyme (adenosylcobalamin, where adenosine is the molecule connected to Co). E. coli obtains Vitamin B12 from LB and it appears that this is what is being delivered to the worm (PMID: 23772381). In CeMM, cobalamin is provided in the form of cyanocobalamin. Whether this is enzymatically converted to cobalamin is not clear. In the model, both cobalamin and adenosylcobalamin are provided from bacterial diet (based on the vitamin B12 content determined in PMID: 23772381) and only cobalamin (representing cyanocobalamin) from other diet (CeMM, or other potential sources in the wild).
Subject: quinone oxidoreductase, 1.6.5.5
1.6.5.5 transfers electrons from nadph to q in cytosol. As a way of regenerating q, demand reaction DMN0033 was added. This demand reaction represents oxidative stress. See PMID: 12435734 for a relevant mammalian enzyme (may not be the same enzyme).
Subject: osmotic pressure, urea, trimethylamine, trimethylamine N-oxide, urea
Tmao (obtained in C. elegans by oxidation of tma, which is predicted to be coming from bacteria) counteracts the denaturing effect of osmolyte urea, and together, tmao and urea are used to establish osmotic pressure in sea animals. A demand reaction was created to address this possible mechanism in C. elegans, which is suggested by PMID: 18077412. Same molar ratio as in sea animals (PMID: 7112124) was used (tmao:urea, 1:2).
Subject: Selenium, selenoproteins
The only selenoprotein in C. elegans is TRXR-1 and it has a single selenocysteine residue (PMID: 21199936). The sources of selenium can be selenite from soil (selanate does not seem to be assimilable by C. elegans due to lack of EC 1.8.4.9) or selenomethionine or selenocysteine from bacterial diet. In the model, selenium is allowed as selenite or bacterial selenocysteine. The latter is ready to be incorporated into trnasec for further processing whereas the former needs to be metabolized through a specific pathway. A demand reaction is inserted to represent the usage of selenocysteine in TRXR-1 synthesis.
Subject: cyanoamino acid metabolism, cyanide detoxification
C. elegans likely encounters cyanide in soil from plants and bacteria (PMID: 21840852 and references therein). Three cyanide detoxification pathways were used: (1) KEGG pathways starting with non-enzymatic reactions and ending in glutamate or alanine, (2) pathway according to PMID: 21840852, which connects to oacetylserine-based detoxification of the h2s initially produced (exists Cysteine and methionine metabolism) and requires uptake of oacetylserine from diet (PMID: 21840852), and (3) KEGG pathway ending in a glutamyl peptide, which is drained by an artificial demand reaction; the fate of this molecule is not clear. The detoxification of cyanide in #2 is connected to H2S detoxification in mitochondria according to PMID: 21840852 and PMID: 19136963. See also note #35.
Subject: h2s, sulfide, h2s detoxification
C. elegans likely encounters H2S in soil from from breakdown of organic sulfur (PMID: 21840852 and references therein). Mitochondrial H2S detoxification system is adapted from PMID: 21840852 and PMID: 19136963. PMID: 21840852 referred to a review (PMID: 20448039), which based their model on PMID: 19136963. In BIGG, there is no mitochondrial h2s, so BIGG human and yeast models missed this pathway. A mitochondrial transport is added since h2s presence in this organelle is implied by this pathway, although the pathway might be detoxifying only h2s produced in the mitochondria. Note that h2s is a small molecule.
Subject: phosphatidylinositol-3-kinase reactions
Phosphatidylinositol-3-kinases of a certain class (class 1) involve a catalytic and a regulatory subunit. Annotations and coexpression data indicate that age-1 and aap-1 are the respective catalytic and regulatory subunits for 2.7.1.153. Although KEGG connects the orthology group of aap-1 to reactions ctalyzed by vps-34 (2.7.1.137) and piki-1 (2.7.1.154) these two connections were removed during first model update because the latter two enzymes do not seem to be class 1 phosphatidylinositol-3-kinases.
Subject: ascaroside biosynthesis
Ascaroside biosynthesis pathways were based on published studies (PMID: 22239548, PMID: 25775534, PMID: 27189572, PMID: 29401383) as well as unpublished data and predictions from the Schroeder lab (https://btiscience.org/schroeder/). Seven abundant ascarosides are made by the current model: ascr1, ascr2, ascr3, ascr5, ascr7, ascr10 and ascr18. Together, these metabolites comprise vast majority of ascarosides produced by C. elegans. Unknown sections of ascaroside biosynthesis, such as the synthesis of ascarylose, were covered by gap filling reactions and pathways that are consistent with mass and charge balance principles and also with similar reactions in the rest of the network. Orthology groups and enzyme identities are not clear for those reactions which are not part of KEGG database.
